Deep-learning atomistic semi-empirical pseudopotential model for nanomaterials

TL;DR

DeepPseudopot is a machine learning-enhanced pseudopotential model that accurately predicts electronic properties of nanomaterials, enabling efficient and transferable simulations for designing new optoelectronic materials.

Contribution

It introduces a neural network-based pseudopotential model trained on GW data, extending traditional methods with improved accuracy and transferability for diverse semiconductors.

Findings

Achieves high accuracy in electronic structure predictions for silicon and III-V semiconductors.

Demonstrates improved efficiency and transferability over traditional pseudopotential methods.

Enables data-driven design of novel nanomaterials for optoelectronic applications.

Abstract

The semi-empirical pseudopotential method (SEPM) has been widely applied to provide computational insights into the electronic structure, photophysics, and charge carrier dynamics of nanoscale materials. We present "DeepPseudopot", a machine-learned atomistic pseudopotential model that extends the SEPM framework by combining a flexible neural network representation of the local pseudopotential with parameterized non-local and spin-orbit coupling terms. Trained on bulk quasiparticle band structures and deformation potentials from GW calculations, the model captures many-body and relativistic effects with very high accuracy across diverse semiconducting materials, as illustrated for silicon and group III-V semiconductors. DeepPseudopot's accuracy, efficiency, and transferability make it well-suited for data-driven in silico design and discovery of novel optoelectronic nanomaterials.